InAs/InAsSb LEDs are a potential solid-state source for mid-infrared gas sensing, where emission can be designed around the characteristic absorption wavelengths of target gases. In this spectral region, carbon dioxide has an absorption feature near 4.2 µm and carbon monoxide near 4.6 µm. The challenge is to produce useful emission at room temperature while limiting the effects of carrier loss and other processes that reduce LED output.
Research into InAs/InAsSb structures has compared strained-layer superlattices (SLSs) with multiple quantum wells (MQWs), examining how composition and layer thickness affect their optical properties. The results offer practical guidance for engineers and device developers assessing semiconductor active regions for mid-infrared emitters.
Why InAs/InAsSb structures suit mid-infrared LEDs
InAs/InAsSb is a type-II semiconductor heterostructure. In a type-II alignment, electrons and holes are concentrated in different parts of the structure, so their recombination is spatially indirect. The band structure can be adjusted through the antimony content of the InAsSb layers and the thicknesses of the constituent layers. This provides a means of tuning the emission energy toward a desired mid-infrared wavelength.
Two forms of the structure are relevant: MQWs, which use quantum well layers separated by barriers, and SLSs, which repeat thin layers to create a superlattice. Structural characterisation using X-ray diffraction and transmission electron microscopy helps assess layer properties and interfaces. Band-structure simulations can then support interpretation of how changes to composition and thickness affect transitions and carrier distribution.
How structure influences emission
Photoluminescence (PL) measurements across 4–300 K showed that increasing the antimony content in the quantum wells shifted emission peaks to longer wavelengths and reduced their intensity. This illustrates a central design trade-off: composition can move emission toward a target gas absorption band, but the optical response must also be evaluated across the device’s intended operating temperatures.
Layer thickness also provides a wavelength-tuning mechanism. Simulations indicated that reducing the thickness of InAs layers raises the energy of the relevant electronic states, allowing the transition energy—and therefore the emission wavelength—to be adjusted. In the structures compared, the SLS showed a greater electron probability within the InAsSb wells than the MQW, increasing the overlap between electron and heavy-hole wavefunctions. That overlap is relevant to radiative recombination and should be considered alongside wavelength when selecting an active-region design.
Temperature, confinement and recombination
Temperature-dependent PL revealed different behaviour in the two structures. At higher temperatures, MQW emission was dominated by electron-hole recombination in the InAs barrier layers. This behaviour was not observed in the SLS samples, where holes remained strongly confined in the quantum wells up to room temperature. Carrier confinement is important because carriers that escape the intended active region are less available for the desired emission.
PL intensity also decreased as temperature rose. The temperature dependence was associated with competing radiative and non-radiative free-carrier recombination. Auger recombination, a non-radiative process in which recombination energy is transferred to another carrier, was investigated in both structure types. The dominant Auger process differed between MQWs and SLSs. Experimental activation energies were larger than theoretical values, indicating some degree of suppression associated with the type-II band alignment.
At low excitation power, the PL spectra also showed excitonic emission: light produced when an electron and hole form a bound pair and recombine. As excitation increased, emission from electron-hole plasma became more prominent. These observations help distinguish emission mechanisms and show why measurements across both temperature and excitation conditions are useful when evaluating an active region.
From photoluminescence to LED electroluminescence
PL characterises optical emission when carriers are generated by light; electroluminescence (EL) measures emission when carriers are injected electrically. Two prototype LEDs incorporating an InAs/InAsSb SLS active region produced mid-infrared EL over a measured 7–300 K range. Their EL spectra showed behaviour similar to the corresponding PL spectra, linking the optical response of the material structure to that of fabricated devices.
At room temperature, the devices produced output powers in the microwatt range. A feature near 4.2 µm, associated with the CO2 absorption fingerprint, was identifiable in the EL spectra. This demonstrates the relevance of the emission wavelength to gas detection, but the reported output also indicates that further device development is needed to improve performance for practical applications.
Design implications for system developers
For engineers evaluating mid-infrared LED sources, the findings point to several considerations:
- Set wavelength through structure design. Antimony composition and layer thickness influence emission, so both should be considered against the target absorption band.
- Assess carrier confinement at operating temperature. The different high-temperature behaviour of MQWs and SLSs shows that low-temperature optical data alone may not predict room-temperature emission.
- Account for non-radiative pathways. Auger recombination and thermal quenching affect the amount of injected carrier energy converted into light.
- Compare material and device measurements. PL can reveal active-region behaviour, while EL confirms how the structure performs under electrical injection.
The comparison identified the SLS as the more attractive of the two investigated structures for an LED active region, particularly because of its carrier confinement and high-temperature emission behaviour. That finding is a design direction, not a guarantee of application-ready output: prototype devices produced microwatt power at room temperature, and further improvement was identified as necessary.
Conclusion
InAs/InAsSb LEDs offer a tunable route to mid-infrared emission relevant to gas sensing. Their performance depends on more than matching an emission peak to a gas absorption feature: composition, layer thickness, carrier overlap, temperature-dependent confinement and non-radiative recombination all shape the result. The SLS findings provide a promising basis for active-region design, while room-temperature device output remains a key practical consideration for system integrators and technical decision-makers.